Hello,
One possible solution would be to get the region of interest by color using the "roicolor" function. http://www.mathworks.com/help/images/ref/roicolor.html
Then using that function you can determine the indices of the image without that color. However, please note that this assumes that this color only appears in the border and not inside the image. For example, consider the following script where I remove the outer white box from an image:
I = imread('sampleImage.png');
I = rgb2gray(I);
colorOfInterest = 255;
roi = ~roicolor(I, colorOfInterest, colorOfInterest);
idx = find(roi==1);
[y, x] = ind2sub(size(I), idx);
x1 = round(min(x));
y1 = round(min(y));
x2 = round(max(x));
y2 = round(max(y));
croppedImage = I(y1:y2, x1:x2);
figure;
subplot(1,2,1);
imshow(I);
subplot(1,2,2);
imshow(croppedImage);
